1. Field of the Invention
The present invention relates generally to stabilizing an optical interface and, more specifically, to creating a reproducible and stable optical interface between biological tissue and an optical blood glucose sensor.
2. Description of the Related Art
Monitoring of blood glucose concentration levels has long been critical to the treatment of diabetes in humans. Current blood glucose monitors involve a chemical reaction between blood serum and a test strip, requiring an invasive extraction of blood via a lancet or pinprick. Small handheld monitors have been developed to enable a patient to perform this procedure anywhere, at any time. But the inconvenience of this procedure—specifically the blood extraction and the use and disposition of test strips—has led to a low level of compliance. Such low compliance can lead to serious medical complications. Thus, a non-invasive method for monitoring blood glucose is needed.
Studies have shown that optical methods can detect small changes in biological tissue scattering related to changes in levels of blood sugar. Although highly complex, a first order approximation of monochromatic light scattered by biological tissue can be described by the following simplified Equation 1:IR=IO exp[−(∥a+μs)L]  Eq. 1where IR is the intensity of light reflected from the skin, IO is the intensity of the light illuminating the skin, μa is the absorption coefficient of the skin at the specific wavelength of light, μs is the scatter coefficient of the skin at the specific wavelength of light, and L is the total path traversed by the light. From this relationship, it can be seen that the intensity of the light decays exponentially as either the absorption or the scattering of the tissue increases.
It is well established that there is a difference in the index of refraction between blood serum/interstitial fluid (blood/IF) and membranes of cells such as blood cells and skin cells. (See, R. C. Weast, ed., CRC Handbook of Chemistry and Physics, 70th ed., (CRC Cleveland, Ohio 1989)). This difference can produce characteristic scattering of transmitted light. Glucose, in its varying forms, is a major constituent of blood/IF. The variation of glucose levels in blood/IF changes its refractive index and thus, the characteristic scattering from blood-perfused tissue. In the near infrared wavelength range (NIR), blood glucose changes the scattering coefficient more than it changes the absorption coefficient. Thus, the optical scattering of the blood/IF and cell mixture varies as the blood glucose level changes. Accordingly, an optical method presents a potential option for non-invasive measurement of blood glucose concentration.
Non-invasive optical techniques being explored for blood glucose application include polarimetry, Raman spectroscopy, near-infrared absorption, scattering spectroscopy, photoacoustics and optoacoustics. Despite significant efforts, these techniques have shortcomings such as low sensitivity, low accuracy (less than current invasive home monitors) and insufficient specificity of glucose concentration measurement within the relevant physiological range (4-30 mM or 72-540 mg/dL). Accordingly, there is a need for an improved method to non-invasively monitor glucose.
Optical coherence tomography, or OCT, is an optical imaging technique using light waves that produces high resolution imagery of biological tissue. OCT creates its images by focusing a beam of light into a medium and interferometrically scanning the depth of a linear succession of spots and measuring the absorption and/or the scattering of the light at different depths in each successive spot. The data is then processed to present an image of the linear cross section of the medium scanned. It has been proposed that OCT might be useful in measuring blood glucose.
One drawback associated with using OCT for monitoring blood glucose is the signal noise associated with optical interferometry, also known as speckle. As discussed in U.S. application Ser. No. 10/916,236 by M. Schurman, et al., entitled “Method and Apparatus for Monitoring Glucose Levels In A Biological Tissue,” to reduce speckle, a glucose monitor incorporating OCT methodology may scan a beam of collimated light continuously and laterally across a two-dimensional surface area of a patient's tissue or skin, while interferometrically scanning the tissue in depth. Preferably, the scanning is accomplished with a small, lightweight, and robust mechanism that can be incorporated into a sensor to be used in a fiber-optics based product or, alternately, a non fiber-optics based product. One main objective of using this type of sensor is to generate a reproducible stable optical interface between the subject's skin and optical path of the sensor in order to take multiple readings from the same lateral location on the skin while maintaining the integrity of the optical interface. As discussed below, there are multiple problems associated with providing and maintaining a stable and reproducible optical interface between an OCT sensor and the skin of a patient.
Two Basic Optic Designs
Two well known sensor designs that use OCT are schematically shown in FIGS. 1 and 2. FIG. 1 shows a design based on the use of two rotating wedge prisms to change the angle of collimated light incident on a focusing lens. In FIG. 1, incoming light beam 101 hits a collimating lens 102, which splits the beam 101 into multiple parallel beams of light, or collimated light 103. The collimated light 103 then passes through one or more wedge prisms 104, which are rotating at predefined rates. As shown in FIG. 1, dual rotating wedge prisms 104 generate an angular deviation in the collimated light 103 from the optical axis of the sensor, which is the “centerline” axis passing through the elements of the sensor, perpendicular to the surface area of skin 109 to be tested. By deviating the angle of the collimated light 103, the focal point of the light moves around on a focal plane of an optical window 108 that is flush against the skin 109, thereby scanning different lateral locations on the skin 109. As shown at 105, once passing through wedge prisms 104, the parallel rays of collimated light 103 may be angled away from the optical axis, depending on what portion of the wedge prisms 104 the collimated light 103 passes through. The angled beams 105 then pass through a focusing lens 106, and begin to focus together to a focal point 107 at the bottom surface of an optical window 108.
FIG. 2 shows a similar concept to FIG. 1, however the dual wedge prisms 104 of FIG. 1 are replaced with an angled mirror 201, for example, a 45 degree angled mirror, that oscillates along two axes, thereby deviating the angle of collimated light 103 from the optical axis in order to move the focal point 108 around on the surface area of skin 109. Accordingly, this OCT sensor design is well known in the art. Both designs facilitate scanning an area of skin by deviating the angle of collimated beam 103 from the optical axis, thereby moving the focal point 107 a proportional distance laterally in the focal plane along the bottom of the optical lens 108, and, accordingly, along the surface area of the patient's skin 109.
While both sensor designs provide mechanisms for incorporating OCT into a noninvasive blood glucose sensor, there are several drawbacks associated with the above designs as described below.
Variations in Optical Path Length
One drawback associated with the dual wedge prism sensor design of FIG. 1 is illustrated in FIG. 3. In an interferometer, the optical path length of a beam of light is determined by the physical or geometric path length of the beam and the index of refraction of the medium which the beam is passing through as shown in Equation 2:LOPT=n·LGEO   Eq. 2where “LOPT” is the optical path length, “n” is the index of refraction, and “LGEO” is the geometric or physical path length.
As shown in FIG. 3, depending on the position of the wedge prisms 104 at the time the collimated beam 103 shines through, while the geometric path length of the collimated beam 103 stays the same, the index of refraction changes due to the changing thickness of the wedge prisms 104 as the prisms rotate, thereby altering the optical path length of the collimated beam 103. This continuous change in the thickness of the wedge prisms 104 continuously alters the optical path length of the collimated beam 103 as it passes through. As shown in FIG. 3, the placement of the wedges may extend the length of the optical path, making it seem as though the skin 109 is moving away from the sensor. Thus, three optical scans taken through the dual wedge prisms 104 when the prisms 104 are in different rotated positions produce three scans beginning at different positions in depth. Since the sensor data is an average of multiple scans, if each scan begins at a different position in depth, the resulting ensemble average will not be representative of a true averaging of multiple scans.
For example, in FIG. 3, when the collimated beam 103 passes through the thinnest area of the wedge prisms 104, as shown at 301, the sensor begins to collect data at Depth A, interpreting the interface between the optical window 108 and the skin 109 to be at Depth A, as shown at 302. However, when the collimated beam 103 passes though a thin portion of the first wedge prism and a thick portion of the second wedge prism, as shown at 302, the sensor begins to collect data at Depth B, interpreting the interface between the optical window 108 and the skin 109 to be at Depth B, as shown at 304. Further, when the collimated beam 103 passes through the thickest portion of both wedge prisms, as shown at 305, the sensor begins to collect data at Depth C, interpreting the interface between the optical window 108 and the skin 109 to be at Depth C, as shown at 306. Since typically multiple scans (e.g., greater than 100 scans) are taken and then averaged to reduce speckle, scans taken at different positions in depth cannot be averaged. Thus, a solution to this problem is desired.
Another drawback associated with the dual wedge prism sensor is the distortion of the scan along the depth axis or z-axis of the light beam entering and exiting the skin. If the rotation speed of the wedge prisms 104 is several orders of magnitude larger than the depth scan rate of the optical sensor, then the depth scale measured by the scan is either “stretched” or “shrunk” by the entire amount of the difference in optical path induced by the changing thickness of the wedge prisms 104. However, if the rotation speed of the wedge prisms 104 is much slower than the depth scan rate, then the changing thickness of the wedge prisms 104 has a minimal effect on the depth scale. For example, if the depth scans occur at 60 Hz, which means that the sensor completes one depth scan within in 1/60th of a second, and the prisms rotate at 3600 rpm, then each wedge prism makes a full rotation during the time it takes the sensor to complete one depth scan. Because the thickness of each wedge prisms varies as the prisms rotate, the optical path length changes during each depth scan, which distorts the depth data collected by the sensor by changing the depth scale during a single scan. Thus, there is an optimization that must occur between the depth scan rate and the prism rotation rate such that the entire surface area is thoroughly scanned while minimizing the z-axis scan distortion.
Scan Pattern Stability
Accordingly, it is desired is that each depth scan be taken at a different lateral position on the surface of the skin 109 such that the ensemble of all the depth scan positions are randomly and uniformly distributed throughout the scan region. The lateral locations of each depth scan must be spatially independent to 1) effectively encompass regions of blood glucose change during a sensor reading and 2) effectively reduce speckle. However, a problem associated with the dual wedge prism sensor in FIG. 1 and the oscillating mirror sensor in FIG. 2 is the inability to capture each depth scan position due to the angular velocity of the wedge prism(s) 104 or the oscillation rate of the angled mirror 201 being harmonic in phase with the depth scan rate of the optical sensor, i.e., the frequency of the angular velocity is a multiple or integral of the depth scan rate of the sensor. When either the angular velocity or oscillation rate is an integral of the depth scan rate, the two rates “beat” against each other, and produce a loss of conformal coverage of the surface area of the skin 109 being scanned.
As shown in FIG. 4B, when using a single rotating wedge prism or oscillating angled mirror in a sensor as described above, the optimal result of multiple depth scans is a circle pattern on the surface area of the skin 109, which each “dot” representing a depth scan. Each depth scan occurs along the path of this circle pattern, effectively breaking the circle up into a series of scanned points. However, if the angular velocity is an integral or harmonic of the depth scan rate, the depths scans begin to overlap in location, thereby producing an incomplete circle pattern and a loss of spatially independent depth scans, as shown in FIG. 4A. With an overlap of depth scans, the same locations of tissue are scanned, causing less speckle reduction and poor imaging of structures within the scanned tissue. The problem becomes even more pronounced in the case of a sensor with two wedge prisms, as shown in FIG. 4C.
Focal Plane Instability
Another challenge presented by both the wedge prism design in FIG. 1 and the oscillating mirror design in FIG. 2 is the inability to maintain the focal point 107 of the focused collimated beam on the focal plane, or the interface between the optical window 108 and the surface area of the skin 109 being scanned. Optical lenses do not project an image onto a flat plane, such as the flat bottom surface of the optical window 108, but, instead, naturally project an image onto a curved surface, much like the curved interior of the eye. This curved surface is well known as a Petzval surface. Thus, as the collimated light 103 enters the focusing lens 106, the focal point 107 of the collimated light 103 traces out a curved focal plane or Petzval surface based on the design of the focusing lens 106, caused by the angular deviation from the optical axis due to the wedge prisms 104 in FIG. 1 or the angled mirror 201 in FIG. 2. Thus, the flat bottom of the optical window 108 does not allow the focal point 107 to remain on the focal plane.
When the focal point 107 moves off of the Petzval surface, the efficiency of the focused light being collected begins to drop, since focal plane is where the light capture is maximized. Additionally, the depth scale of the focused light is affected such that the displacement of the focal point 107 off of the focal plane results in an equivalent loss in the depth scale of the signal. This results in a blurring of the optical axis, causing measurable details within the skin to be blurred or washed out. Thus, a displacement of the focal point off the focal plane results in a reduction in the sensor signal intensity and a blurring of the optical axis.
Additionally, optical lenses are not perfect. Therefore, as the focal point 107 moves away from the optical axis due to the rotating wedge prisms 104 or the oscillating angled mirror 201, the focused beam drifts away from the skin 109 and back towards the focusing lens 106, and, thus, moves off the focal plane. As discussed above, when the focal point 107 is no longer on the focal plane, the collection efficiency of the light drops, resulting in the collected data incorrectly indicating a reduction in power. This, in turn, alters the depth of the focused beam, thereby unwittingly washing out details in the skin and lowering the resolution and integrity of the scan.
Skin/Sensor Optical Interface
The surface of the skin is “rough” relative to the light entering and exiting the skin during an optical scan. This is well known as optical roughness. Additionally, the refractive index of the skin being scanned typically is different from the refractive index of the material of an optical window of a sensor. As shown in FIG. 5A, the optical window 503 is not necessarily flush against the surface of the skin 504, due to optical roughness 505 of the skin. Accordingly, as incident light 501 is directed towards the skin, some of the light is reflected and/or diffracted, as shown at 502, because there is a mismatch between the index of refraction of the optical window 503 and the index of refraction of the skin 504. This mismatch of refractive indices and, in addition, the space between the skin 504 and the optical window 503 due to the optical roughness 505 reduces the reliability of data taken by the sensor.
FIG. 5B displays two scans taken at the same location on the skin but measured at different points in time with constant optical contact between the skin 109 and the optical window 108 of a sensor. Such scans may be produced by either the dual wedge prism sensor of FIG. 1 or the angled mirror sensor of FIG. 2. Data line 506 represents an averaged optical scan taken at Time 0 while data line 507 represents an averaged optical scan taken thirty minutes after Time 0. Typically, the focused beam hits the interface between the optical window 503 and the skin 504, a sharp rise or peak in the signal is produced, as shown at peaks 510 and 511. The signal then drops as the beam moves through the skin 504 and begins to rise again as the beam hits the interface between the epidermis and dermis layers, as shown at peaks 508 and 509. The signal again drops and continues to drop as the beam reaches the desired depth then returns back to the sensor.
As shown in FIG. 5B, while constant optical contact is maintained between the skin 504 and the optical window 503 of the sensor, over time the optical signal drifts, as illustrated by the peaks at the interface between the dermis and epidermis layers, which rises over time, from peak 508 at Time 0 to peak 509 at Time 0+30 minutes. However, the peak at the interface between the optical window 503 and the skin 504 drops over time, from peak 510 at Time 0 to peak 511 at Time 0+30 minutes. This change in signal intensity is due to a gradual change in the optical interface created by an accumulation of sweat and skin oils at the interface of the optical window 503 and the skin 504, as shown at 512 in FIG. 5C, which serves as an optical transition for the incident light 501 to efficiently travel from the optical window 503 to the skin 504. Additionally, the accumulation of sweat and skin oils smoothes out the optical roughness of the skin. Although the refractive index between the optical window 503 and the skin 504 will stabilize or reach an equilibrium value due to sweat, oil, and other fluids produced by the skin over time, this process could take upwards of 60-90 minutes. Unfortunately, these changes in signal intensity over this extended period of time may completely mask the changes that are occurring along the OCT signal, and thus prevent proper correlation of changes in the OCT signal to changing glucose levels, as discussed in U.S. Provisional Applications Nos. 60/671,007 and 60/671,285, both entitled “Method For Data Reduction and Calibration of an OCT-Based Blood Glucose Monitor.” Thus, multiple scans taken over time cannot produce a reliable measurement from the same lateral location on the skin. In addition, a patient would be required to place the sensor onto his or her skin and wait 60-90 minutes before using it, in order to receive reliable and reproducible results, which creates an inefficient sensor.
Thus, a need exists for an optical sensor for measuring blood glucose levels and other physiological effects that overcomes the deficiencies discussed above.